Chapter 2 Behavioral Variability and Research 1 Behavioral

Chapter 2: Behavioral Variability and Research 1. Behavioral science involves the study of variability in behavior • how and why does behavior vary across situations, people, and time. 2. All research questions in behavioral sciences are about behavioral variability.

3. Research should be designed to permit the researchers to describe and account for the variability in behavior. • A good study will determine why behavior varied. 4. Measurement of behavior involves the assessment of behavioral variability. • A participant’s score should correspond to the variability in a his or her behavior.

5. Stats are used to describe and account for observed variability in behavior. • Descriptive statistics: describe behavior, using numbers like averages and percentages. • Inferential statistics: used to make conclusions about significance and reliability and generalizability of results.

Variance • Depicts the amount of variability in participants’ behavior. The dispersion of scores. Range: difference between largest and smallest score > but range does not take into account the other scores in your sample. Usually variability is described in terms of how much the scores vary around the mean (average).

Low Variability

High Variability

1. Standard deviation (s) is the square root of the variance (s 2). 2. Thus, the larger the variance, the larger the standard deviation, and the more variability (dispersion) in the scores.

Systematic and Error Variance • • Systematic variance: part of total variance in behavior that is related in an orderly, predictable way to the variables you are researching. > Aggression scores may vary systematically to changes in TV violence viewing. Error variance: variance that is not accounted for or unrelated to variables the researcher measures. Factors that the researcher does not study that may relate to variability in behavior.

Total variance = systematic variance + error variance

Strength of Relations Proportion of total variance that is systematic: • systematic variance/total variance Effect Sizes: measures of the strength of association between two variables • range from 0. 0 to 1. 00 • 0: none of variability is systematic (0/100) • 1: all of the variability is systematic (100/100) • The larger the proportion the stronger the relation between the two variables

Strength of association between two variables: • small. 01 • medium. 06 • large. 15+ • Thus, a strong association between two variables can account for only 15% of the total variance, leaving 85% error or unaccounted for variance.

Meta-Analysis • One study will provide a rough estimate of of the strength of association between two variables • But, remember the importance of replication! Meta-Analysis: statistical procedure used to analyze and integrate the effect sizes from many different studies to get a single estimate of the strength of association between two variables. • Also used to explore other factors that may be affect the relationships.

Chapter 3: The Measurement of Behavior Types of Measures 1. Observational measures: direct observation of behavior > time to complete task, number of bar presses 2. Physiological measures: internal processes that are not observable > heart rate, brain waves

3. Self-report measures: participant gives answer Cognitive: what people think about something > child’s response when answering a question such as which ball is heavier Affective: participants’ report of how they feel > depression questionnaire Behavioral: participants’ report of how they act > questionnaire on how child behaves in school

Measurement To assign numbers to a participants’ response (measured variable) so it can be analyzed and correspond in a meaningful way to what the variable you a are measuring (conceptual variable). 1. Measured variables: numbers that represent measured variables 2. Conceptual variables: what you are try to measures (anxiety, self-esteem) An operational definition is a statement of how a conceptual variable is turned into a measured variable.

Psychometrics: the study of psychological measurement in order to improve measures and reduce error. Converging operations: using different measures for a single concept (e. g. anxiety) to get a better measure of the concept (triangulation).

Scales of Measurement 1. Nominal Scale: a number (label) is assigned to name or identify a particular characteristic. 1. 1 blue eyes, 2 brown eyes, 3 green eyes etc. 2. Ordinal Scale: rank order the variable. Numbers indicate if there is more or less of a variable but not the distance between the variables. > Score of 5 is better than 1, but does not say how much better it is.

3. Interval Scale: equal differences between points on the scale correspond to equal differences in characteristic measured. > • IQ test, difference between 85 and 95 (10) is same as between 105 and 115 (10). Does not have a true zero, so cannot multiply or divide numbers. 4. Ratio Scale: has a true zero point so numbers can be multiplied and divided. > Weight: 100 pounds is twice a heavy as 50.

Reliability • The consistency and stability of a measure. The extend to which it is free from error. > IQ is 100 one day and 140 next day, then the measures is unreliable. Observed score = true score + measurement error True score: real score if there is no error Measurement error: factors that distort the score

Measurement Error Measurement error may be due to: • transient states of participant (mood, fatigue) • stable attributes of participant (motivation, IQ) • situational factors (lighting, harsh researcher) • measure characteristics (ambiguous questions, long test) • mistakes in recording responses

Assessing Reliability Total variance in set of scores = variance due to true scores + variance due to measurement error. • Variance due to true scores is systematic variance Reliability = True score variance/Total variance • proportion of total variance in a set of scores that is systematic (true score) variance.

• 0 = none of total variance is true score variance • 1 = all of total variance is true score variance • Reliability ranges from 0. 0 (no reliability) to 1 (perfect reliability). Reliability of. 70 is considered good meaning 70% of total variance is true score (systematic) variance.

Test-retest reliability: The extent to which scores on the same measure, administered at two different times, correlate with each other. Interitem reliability: The extent to which scores on items of a scale correlate with each other. • Item-total correlation: correlation between one item and the sum of all the other items. • Split-half reliability: correlation of scores on one half of test with scores on another half. > Cronbach’s alpha

Equivalent-forms reliability: The extent to which scores on similar, but not identical measures, administered at two different times, correlate with each other. > Does IQ test A correlate with IQ test B? Interrater reliability: The extent to which ratings of one or more judges correlate with each other.

Increasing Reliability • Standardize test administration • Use clear instructions and questions • Train observers • Minimize errors in coding data